The world’s first symmetric multi-processor (SMP) to be powered by Intel’s Many Integrated Core (MIC) architecture is also the latest incarnation of Stephen Hawking’s pioneering Cosmos supercomputer for the Miracle Consortia. Cosmos makes use of SGI’s “Big Brain”–the SGI UV 2000–which houses 32 Xeon Phi co-processors utilizing a total of 1,856 cores in a cache-coherent shared-memory space of nearly 15 terabytes (TB).
“With our powerful and flexible SGI UV 2000 we can continue to focus on scientific discovery, leading worldwide efforts to advance our understanding of the universe,” says Stephen Hawking, founder of Cosmos and the Miracle Consortia. “Our newest Cosmos supercomputer from SGI contains the latest many-integrated core technology from Intel, the Xeon Phi coprocessor.”
The Cosmos supercomputer and the Miracle Consortia pursue the most vexing problems in cosmology. Consortia members include top British universities (see full list below).
SGI’s UV 2000–nicknamed the Big Brain–powers Stephen Hawking’s Miracle Consortia’s Cosmos supercomputer, which houses 32 Xeon Phi coprocessors in the world’s first MIC-powered SMP. Source: SGI
Phi-Based Supercomputer Tackles Big Cosmic Questions
The latest COSMOS supercomputer will support key UK research in several key areas:
- Discovering new extra-solar planets
- Assist in lattice field-theory simulations of the early universe
- Validate new cosmological models–including superstrings and extra dimensions
- Study cosmic background radiation, galaxy distributions and the generation of primordial gravitational waves.
- Answering “big brain” questions such as “how many exoplanets have the right conditions for life?”
SGI’s UV 2000 uses NUMAlink blade-based architecture to share a single system image (SSI) among up to 2,048 cores (4,096 threads), enabling vast data sets to tackle the world’s most complex simulations. NUMAlink supports 50 giga-byte per second data transfers–in four 12.5 Gb/s channels–while allowing a global coherent shared memory scalable to 32,000 sockets. And the next generation of SGI’s UV promises to expand support for up to 4096 Xeon Phi cores 8192 threads) sharing 64-TB of cache-coherent shared memory in a SSI.
“The SGI UV 2000, when paired with Intel Xeon Phi coprocessors, offers in-memory computing at an extreme scale,” explains SGI CTO Eng Lim Goh. “COSMOS will be able to accelerate finding answers to some of the world’s most data-intensive problems using software tools with which they are already familiar.”
SGI also recently updated its entire line of supercomputer-caliber servers for use with Intel’s MIC architecture, offering Xeon Phi coprocessors in factory-integrated hardware and software “starter kits” for the SGI UV 20 with four Intel Xeon E5-4600 processors and four PCIe slots for Xeon Phi coprocessors, the SGI Rackable twin-socket Intel Xeon server with four Xeon Phi slots, and the SGI ICE X, the cale-out blade server using PCIe-capable InfiniBand connectivity.
MIC-Powered Supers Continue Spread Worldwide
Besides the COSMOS Consortium, MIC architecture Xeon Phi powered SGI supercomputers are already in use at NASA (Ames) as well as at The Genome Analysis Centre (TGAC, Norwich, U.K.)
Cosmos is a U.K. Science and Technology Foundation Council (STFC)-funded DiRAC Consortia project to extend England’s high-performance computing (HPC) capabilities, and is part of the U.K.’s $253 million effort to improve its e-infrastructure.
Members include the University of Cambridge, the University of Sussex, Portsmouth University, the University of Manchester, Imperial College London, the University of Central Lancashire, the University of Nottingham, the University College London, the University of Oxford, and the University of Durham in the U.K. plus Eotvos Lorund University in Hungary.
i am somewhat surprised that the people in some countries are so taken by prof hawking , rather than the person he worked with many years ago , prof r penrose
@perih60 I did not say that I was taken with Professor Hawking. it is great to see him surmounting his physical shortcomings and being able to continue his theoretical career. We must remember that so far most of his M-theory is still not a unified theory. Professor Penrose and others take issue with his claim to this M-theory as a proven theory, but just a more advanced string theory. At least that is what I have read. I do not think that we will see a proven theory in our day. After all most people forget that Darwin's theory of evolution is just that - a theory. I have yet to see an ape, chimpanzee or monkey in the throes of becoming a human person - or even finding the so-called 'Missing Link!' It is only when you can actually prove something without any doubt that a theory is proven, in my view. This is why I am not certain that M-theory has been proven to be true.
@Suizi Your mistaking theory with hypotheses. Theory is "a well-substantiated explanation of some aspect of the natural world, based on a body of facts that have been repeatedly confirmed through observation and experiment." Hypotheses is unproven. If you're not believing in evolution I'm sure you'd have no problems taking just penicillin as an antibiotic, after all, the bacteria couldn't possibly evolve to be resistant to it.
@Tmt Thank you @Tmt. This was what I was trying to say. I just didn't use the right terminology. I have seen no convincing evidence of Darwin's 'Theory' of evolution, so I believe this is just an hypothesis. I have seen no convincing evidence that M-theory has been proved, therefore I regard it, too, as an hypothesis, too. i don't even believe it is a fully functional working hypothesis, either.
Aaah! But does this Big Brain solve the Travelling Salesman problem? And if so, how quickly?
@Suizi saling is wrong
down with zee moneitary system
when zee computers will rule will will have no need 4 any curency
zeitgeistely
@tontonremi @Suizi sure and BB is watching
@RikDeedit @tontonremi @Suizi I know that BB is watching - and has been doing so from the moment I was born. BB is too big a problem for me to solve. It would be helpful if everyone worked together to de-fang BB. But that is most unlikely to happen - so I will live my life the way I choose; believe what I choose to believe and change the things that I can change. I am too old to do anything else, now, but that does not mean I am unaware of many things that are happening in our world at this time. I just choose to enjoy life and not get paranoid and worried over the things that I cannot change without the help of many, many others. that may seem like a cop-out but that is how I feel. :)
@tontonremi @Suizi I do not understand your point - or if indeed you are actually making one. I am sorry that you seem to be hooked up to some ideological/philosophical theory that may or may not be true but does not really have anything to do with the subject of this 'Big Brain" computer. I was interested in finding out how long this computer would take to give a definitive answer to a logical question often posed to computers to assess their capability. Nothing whatsoever to do with zeitgeist which has to do with the spirit, attitude or general outlook of a specific period, especially as it can be reflected in literature or philosophy for instance. We are talking about the computing capability of this SGI UV 2000 computer.
zee zeitgeist movement which advocates cybernetics is following this matter
zeitgeistely